Project January 2008

Transcription

1 IMPLEMENTATION OF GPS CONTROLLED HIGHWAY CONSTRUCTION EQUIPMENT PHASE II Project January 2008 National Center for Freight & Infrastructure Research & Education College of Engineering Department of Civil and Environmental Engineering University of Wisconsin, Madison Authors: Alan Vonderohe University of Wisconsin, Madison Principal Investigator: Alan Vonderohe Professor Emeritus, UW-Madison

2

3 Implementation of GPS Controlled Highway Construction Equipment Phase II WisDOT Project ID: CMSC: 2006 WO 1.8 Final Report January, 2008 Submitted to the Wisconsin Department of Transportation Alan P. Vonderohe Construction and Materials Support Center University of Wisconsin Madison Department of Civil and Environmental Engineering

4 DISCLAIMER This research was funded by the Wisconsin Department of Transportation and the National Center for Freight and Infrastructure Research and Education. The contents of this report reflect the views of the authors, who are responsible for the facts and the accuracy of the information presented herein. This document is disseminated under the sponsorship of the Department of Transportation, University Transportation Centers Program, in the interest of information exchange. The U.S. Government assumes no liability for the contents or use thereof. The contents do not necessarily reflect the official views of the National Center for Freight and Infrastructure Research and Education, the University of Wisconsin, the Wisconsin Department of Transportation, or the USDOT s RITA at the time of publication. The United States Government assumes no liability for its contents or use thereof. This report does not constitute a standard, specification, or regulation. The United States Government does not endorse products or manufacturers. Trade and manufacturers names appear in this report only because they are considered essential to the object of the document.

5 Final Report Implementation of GPS Controlled Highway Construction Equipment Phase II Alan P. Vonderohe Construction Materials and Support Center University of Wisconsin Madison January, 2008

6 Technical Report Documentation Page 1. Report No. CFIRE Title and Subtitle 2. Government Accession No. 3. Recipient s Catalog No. CFDA Report Date January, 2008 Implementation of GPS Controlled Highway Construction Equipment - Phase II 6. Performing Organization Code 7. Author/s Alan Vonderohe, UW-Madison 9. Performing Organization Name and Address Construction and Materials Support Center University of Wisconsin-Madison Department of Civil and Environmental Engineering -and- National Center for Freight and Infrastructure Research and Education (CFIRE) University of Wisconsin-Madison 1415 Engineering Drive, 2205 EH Madison, WI Sponsoring Organization Name and Address Wisconsin Department of Transportation 4802 Sheboygan Avenue Madison, WI Performing Organization Report No. CFIRE Work Unit No. (TRAIS) 11. Contract or Grant No. WisDOT: Type of Report and Period Covered Final Report [5/1/07 4/1/08] 14. Sponsoring Agency Code 15. Supplementary Notes Project completed for the CFIRE with support from the Wisconsin Department of Transportation. 16. Abstract During 2006, WisDOT and the Construction Materials and Support Center at UW-Madison worked together to develop a specification and QC/QA procedures for GPS machine guidance on highway construction grading operations. These specifications and procedures are intended for incorporation in contracts on two to five pilot projects during the 2007 construction season. The 2006 work, and the 2007 pilot projects are the first two steps in a phased implementation plan that includes refinement of the specification and procedures after the 2007 pilots, additional pilots during 2008, and potential statewide implementation of optional GPS machine guidance for grading on 2009 contracts. To be effective, the goals, objectives, and methods for data collection and analysis for the 2007 pilot projects must be well-articulated; the project sites must be carefully selected; and a set of deliverables and management plan for coordination of the effort must be developed. 17. Key Words GPS, machine guidance, construction, implementation 18. Distribution Statement No restrictions. This report is available through the Transportation Research Information Services of the National Transportation Library. 19. Security Classification (of this report) Unclassified 20. Security Classification (of this page) Unclassified 21. No. Of Pages Price -0- Form DOT F (8-72) Reproduction of form and completed page is authorized.

7 Table of Contents Section Page 1. Background 1 2. Specifications and QC/QA Procedures 1 3. Issues Intended to be Addressed by the Pilot Projects 1 4. Information Gathering Methods 2 5. Northeast Region (Green Bay) Projects Discussion Points Raised by the Region Pilot Project (STH 57) Discussion Points Raised on the Project Project (STH 57) Discussion Points Raised on the Project Southwest Region (Madison) Pilot Project Pilot Project (STH 106) Discussion Points Raised on the Project Summary of Issues for Discussion Lessons Learned from 2007 Pilots Revisions to Specification Revisions to Guidance Language Summary of Issues Intended to be Addressed by the 2007 Pilot Projects Plans for Potential 2008 Pilot Projects Issues to be Addressed during 2008 Pilot Projects Outreach The Bigger Picture A Period of Transition 25 Acknowledgements 26 i

8 List of References 26 Appendix A. Specification and Guidance Language for 2007 Pilot Projects (As Presented in Vonderohe (2007a)) 27 Appendix B. WisDOT Southwest Region s Revision to Specification 40 Appendix C. Record-Keeping Forms and Examples Distributed to Pilot Project Engineers and Foremen 42 Appendix D. Questions and Talking Points for Second Site Visit to Hoffman Project 45 Appendix E. Questions and Talking Points for Second Site Visit to Wondra Project 49 Appendix F. GPS Work Plan Submitted by Hoffman Construction 53 Appendix G. Site Calibration / Localization Checks for Hoffman Project 59 Appendix H. Subgrade Checks for Hoffman Project 62 Appendix I. GPS Work Plan Submitted by Wondra Excavating 68 Appendix J. Site Calibration / Localization Checks for Wondra Project 72 Appendix K. Subgrade Checks for Wondra Project 76 Appendix L. Specification for 2008 Pilot Projects 80 Appendix M. Guidance Language for 2008 Pilot Projects 88 ii

9 1. Background During 2006, WisDOT and the Construction Materials and Support Center (CMSC) at UW-Madison worked together to develop a specification and QC/QA procedures (i.e., guidance language) for GPS machine guidance on highway construction grading operations. These specifications and procedures were incorporated as change orders in contracts on two pilot projects during the 2007 construction season. GPS machine guidance was also evaluated on a third 2007 construction project. Useful information was obtained from the third project, although it was not officially designated as a pilot. One of the pilot projects was in WisDOT s Southwest Region (Wondra Excavating was the contractor) and the other was in WisDOT s Northeast Region (Hoffman Construction was the contractor). The third project was also in the Northeast Region (Mashuda Contractors was the contractor). The 2006 work and the 2007 pilot projects were the first two steps in a phased implementation plan that includes refinement of the specification and procedures after the 2007 pilots, additional pilots during 2008, and, ultimately, statewide implementation of optional GPS machine guidance for grading. CMSC assisted WisDOT with evaluation of the specification and guidance language on the two pilot projects, with the goal of identifying and making any necessary modifications in a timely manner, so that revised versions were in place for bidding on the 2008 pilots. 2. Specifications and QC/QA Procedures The Northeast Region (Hoffman) pilot project adopted the specification and guidance language as they appear in Vonderohe (2007a) and in Appendix A of this document. The Southwest Region (Wondra) pilot project incorporated a modified version of the specification (see Appendix B), but used the same guidance language. The primary difference between the two specifications is that the Southwest Region s required bluetop stakes on the first 3000 feet of subgrade and the Northeast Region s did not require any blue-top stakes. 3. Issues Intended to be Addressed by the Pilot Projects General questions addressed on the pilot projects include: Equipment: 1. What are the frequency, duration, and types of problems with operation of the technology (e.g., poor satellite geometry, loss of lock, multipath, software glitches, data entry and other human errors, technology incompatibilities)? 2. What are the vertical tolerances that are achievable using GPS Machine guidance? 3. What are the necessary knowledge and skill levels for project engineers, contractor project managers, and machine operators? 4. What other efficiencies are realized with GPS machine guidance? 5. What other difficulties arose with GPS machine guidance? 6. Is it necessary for the height modernization program to be completed in the area for GPS machine guidance grading to be successful? If not, what 1

10 additional project control is needed and what other challenges are encountered? Department Responsibilities: 1. Are 3D model components provided by WisDOT readily usable by contractors? 2. What are the frequency and causes of revisions to three-dimensional models? 3. Are data exchange standards and rates sufficient for updating models during construction? 4. What is the appropriate spatial frequency for quality assurance checks and what are the appropriate tolerances? 5. Are there issues, during construction, with utility coordination, subcontractors, or others due to reduced staking requirements? Contractor Responsibilities: 1. What is the appropriate control configuration for GPS site calibration? 2. What are the appropriate tolerances for GPS site calibration? 3. What is the appropriate frequency for GPS site calibration checking / recalibration? 4. What is the maximum geographic extent over which a single GPS site calibration is valid? 5. What is the appropriate spatial frequency for quality control checks and what are the appropriate tolerances? 6. What needs to be staked and what staking can be eliminated? 7. What is the appropriate number of supplemental control points to be furnished by the contractor and what are the appropriate methods for development of supplemental control? 8. Do minor field changes need to be incorporated in the 3D model or is it more cost effective to have the machine operator override automatic controls for minor changes? 9. What types of field changes warrant changes in the 3D model? 4. Information Gathering Methods Information from the pilot projects was gathered by the following means: 1. Project plans and schedules. 2. Telephone and face-to-face interviews with project personnel (both WisDOT region staff and contractor staff). 3. Attendance at pre-construction and weekly progress meetings. 4. Two rounds of site visits to the pilot projects and a site visit to the third project. The first site visit to the Wondra project was conducted on May 23, The second site visit to the Wondra project is scheduled for August 22, The first site visit to the Hoffman project was conducted June 19-20, The second site visit to the Hoffman project was conducted on August 1, The Mashuda project was also visited on August 1, Acquisition of project data, including 3D models, GPS work plans, project engineer diary entries, contractor reports, and completed spreadsheets 2

11 containing measurements. The spreadsheet templates (see Appendix C) were developed by the author and distributed to project engineers and foremen. 6. Sets of questions and talking points, distributed to region staff and project personnel prior to the second site visit to each pilot project (see Appendices D and E). These were used to generate discussion during interviews. The questions and talking points differed for the two pilot projects because the projects operated under different specifications. 5. Northeast Region (Green Bay) Projects Region Project Development Supervisor: Steven Noel Region Data Manager for Pilot Projects: Brad Hollister Region Project Manager for Pilot Projects: Dan Segerstrom Region Design-Side Survey Coordinator: Mike Vandehei Region Construction-Side Survey Coordinator: Dennis Keyzer The Northeast Region s GPS machine guidance pilot project (Hoffman) was along STH 57 in Door County. The second project (Mashuda) in the Northeast Region was also along STH 57, adjacent to and north of the officially-designated pilot project. The region developed breakline data from plans to provide to contractors on both these projects. By mid-may, the region had devoted approximately 40 person-hours to data development for each of the two projects, including mainline and side roads. Data can be developed along the mainline at a rate of approximately four hours per mile, but side roads are developed much more slowly. Side roads require more attention, in part, because match lines are sometimes done incorrectly with no closure between adjacent roadways or with slopes that do not intersect. As of August, no requests had been received by the region to modify any design because of flaws in the plans. However, there were numerous issues with formats for data exchange. The LandXML format in the pilot project specification is not really a standard. Various software vendors have their own versions of this format. Therefore, LandXML could not be used as an exchange format for either the Hoffman or Mashuda projects. Through testing, it was determined that Autodesk s.dwg format provided the most complete and useful data exchange. Breakline data provided by WisDOT is not the final design surface model needed for GPS machine guidance. Development of the design models from the breaklines and other plan information is the responsibility of the contractor under the 2007 specification. The Methods Development Unit (MDU) checked the 3D model developed by the contractor. Procedures developed for processing the data were fully documented. The region was interested in determining if GPS rover data can be used for pay measurement purposes. They were also interested in exploring possibilities for integration of WisDOT s roadway design system and WisDOT s field documentation system. The Northeast Region has a design-side survey coordinator and a construction-side survey coordinator. Design-side surveying begins with aerial mapping, runs through right-of-way layout, and includes any survey work needed for design until the time of construction. It includes WisDOT-provided geodetic control. Construction-side surveying includes fulfilling project manager requests for surveys, borrow pits, and final details for quantities. Surveying operations use both in-house crews and consultants. 3

12 Geodetic control for both the Hoffman and Mashuda projects was established by a single consultant using rapid static GPS methods. This control consisted of 20 monumented points, well-distributed along a ten-mile corridor of the highway. Leveling was done for one-third of the monuments to check GPS-derived orthometric heights. The consultant s geodetic control was not tied directly to the HARN or to any height modernization control points. Rather, the control was tied to section corners whose coordinates had been determined ten years previously using rapid static GPS methods tied to the HARN. Vertical control points were tied to NGS benchmarks by a ten-year-old WisDOT survey. The consultant found a discrepancy in one of the WisDOT vertical control lines and distributed the error linearly along the line. The horizontal datum for geodetic control was NAD 83 (1991) and the vertical datum was NGVD 29. Horizontal positions were expressed in Door County coordinates. Construction-side surveys are tied to known coordinates and elevations of right-of-way monuments, using total stations and levels. To compute final quantities, an as-built DTM is developed from detail survey data. Cross-sections are then cut through the DTM. The average-end-area method is then used to compute quantities from the cross-sections. The first site visit to the Hoffman project was held in conjunction with an open house, on GPS machine guidance, hosted by the Northeast Region. There were attendees Among presentations at the open house was one given by the author. WisDOT s Rick Larson also gave a presentation on issues with 3D models and data exchange. The open house included a field trip to the Hoffman project site Discussion Points Raised by the Region For consideration for the 2008 specification and beyond, region personnel raised the following issues: 1. What will be WisDOT s commitment on provision of design-side data? Considerations include limited resources, the period of transition from CAiCE to Civil 3D, allotted time for making changes (e.g., some changes require more time than others), and compensable versus non-compensable changes in the data. 2. What is the best means for getting other WisDOT regions involved? 3. What effective educational mechanisms are there for WisDOT and contractor staff? 4. Will future GPS machine guidance grading specifications include slopes and ditches or will they continue to be confined to roadway subgrade? 5. When WisDOT goes to statewide options for bidding, will the regions have choices or will the option apply to all contracts? 6. Is it possible to use existing section corners for horizontal control? 4

13 5.2. Pilot Project (STH 57) WisDOT Northeast Region Field Engineer: Barry Paye Contractor: Hoffman Construction Contractor Project Manager: Chris Goss Contractor Project Foreman: Frank Laufenberg Contractor Grading Foreman: Ken Bork This project involved improving 5.2 miles of STH 57 from bi-directional to divided highway in southern Door County. The project ran from just south of the county line at station to Truway Road at station Considerable earthwork was included, with rock cuts as deep as 39 feet and fills as high as 25 feet. GPS machine guidance was used on the full length of the project. Hoffman has been using GPS machine guidance for the past three years and has a number of dozers and graders outfitted with Trimble technology (see Figure 1). The grading foreman on the pilot project was quite familiar with GPS machine guidance and serves as an instructor on use of the technology for machine operators statewide. Hoffman subcontracted for staking on the project. Figure 1. Hoffman Grader Mounted with Trimble GPS Machine Guidance NOTE: Two antennae mounted above the top of the blade give the blade coordinates, direction and roll angle. The pitch angle is computed, by the on-board software, from immediately-preceding blade positions. The field engineer had experience with GPS machine guidance for both grading and base course, having worked with MinnDOT on field verification during its first GPS machine guidance project. Work began in early April, with the GPS machine guidance equipment arriving on site during the second week of May. Hoffman submitted their GPS work plan (see Appendix F) and conducted work accordingly. Initial site calibration was 5

14 completed (see Figures 2 and 3) and the field foreman provided weekly calibration files and calibration check data to the field engineer. According to specification, control points not used in the initial calibration were used to check that calibration. Calibration checks were consistently within the specified tolerances of 0.10 ft horizontally and 0.05 ft. vertically, although one control point was found to be deficient and was not used. Hoffman did not have to densify the WisDOT-provided control. Appendix G contains the project s site calibration check values. The Door County terrain is such that the base station radio range was approximately 1.5 miles. This could be boosted to three miles with a repeater. So, two base station locations were needed to cover the 5.2-mile job. After leaf-out in late May, there were some minor problems with satellite signal reception on steep back slopes and there are sometimes minute periods of poor satellite geometry. The field engineer kept an electronic diary, including entries pertinent to GPS machine guidance aspects of the job. Figure 2. Hoffman s Base Station (a) (b) Figure 3. Ken Bork Performing a Calibration Check (a) and Calibration Check Results (b) 6

15 When breaklines, provided by WisDOT in LANDXML format, were imported into Trimble s Terramodel software, measurement units became confused. The software expected international feet and the data were in US survey feet. Also, LANDXML does not import either planimetric attributes or colors for line work. These problems do not exist if.dwg (AutoDesk) format is used for data exchange. Figure 4 provides two views of segments of the project data after the 3D model was constructed. a. Breaklines and Stationing b. Plan View of Surface Model Figure 4. Two Views of Segments of the Hoffman Data 7

16 The field engineer and region staff were not able to view Hoffman s 3D model directly because they did not have Terramodel. However, the region is supposed to check the 3D model. To facilitate this, Hoffman exported breaklines from Terramodel, and region staff imported those breaklines into CAiCE. CAiCE was then used to build a surface model from the breaklines. There was no guarantee that this was the same surface that Hoffman used because there are different algorithms for inserting breaklines in surfaces and it is not known if the two programs use the same algorithm. A further complication was that designs, prepared by consulting firms not using CAiCE, had to be compared on paper to slices through the CAiCE model. For the required minimum of 20 subgrade checks per mile, Hoffman decided to identify random points and then move from each of those points to its nearest cross-section so that checks could be made directly against the plans and not against the design surface that was generated from the plans. Subgrade checks along the mainline began in August. The complete set of 230 subgrade check values appears in Appendix H. The largest positive difference is ft. The largest negative difference is ft. The mean of the check values is ft and the standard deviation is ±0.06 ft. The 2007 specification required the contractor to inform the engineer if more than one of any five consecutive subgrade checks exceeded 0.10 ft. in absolute value. In only three instances was this specification exceeded: 1. At Sta R, the check value was ft. At Sta , the check value was ft. These were two consecutive checks made at nearly the same location. 2. At Sta L, the check value was -0.12ft. At Sta R, the check value was ft. These two check points had three other check points between them with all three having absolute values less than 0.10 ft. The two offending check points were physically separated by 700 ft. 3. At Sta L, the check value was ft. At Sta L, the check value was ft. These two check points had two other check points between them, each of which had an absolute value less than 0.10 ft. No blue top stakes were used on the Hoffman project. Most of the slope and ditch grading on the Hoffman project was done using GPS machine guidance. Project engineers and foremen reported more-than-adequate results, with visually smooth linear features and surfaces (see Figure 5) Discussion Points Raised on the Project 1. What is the best way to coordinate design changes with 3D model changes and who does what? What if the model is built by a third party (this was the case for the Mashuda project)? 2. The project engineer and region staff could not even view the model without the vendor s software. 8

17 Figure 5. Slopes and Ditches Constructed with GPS Machine Guidance on Hoffman Project 3. Some minor design changes might not be incorporated in the 3D model, so it cannot be used for as-builts. 4. There are problems with using last-pass data from the machines to construct as-built surfaces. For example, data are collected whether the blade is cutting, flush with the ground, or above ground. Also, such data can be collected only as points. No breaklines or labels can be inserted, so the data would have to be interpreted and edited later with no information other than visual inspection on which points should be connected to form lines. 5. The 2007 specification allowed the engineer to require conventional staking if GPS machine guidance was producing unacceptable results. Should the contractor also have the option of going to conventional staking (e.g., equipment priorities)? 6. GPS work plan, site calibration check, and subgrade check requirements were good in the 2007 specification. Hoffman had no problem meeting the specified tolerances. The belief was that those tolerances are appropriate and should not be made more stringent. 7. Why is it necessary for the site calibration check points to be different from the control points used for calibration? If the site calibration checks once, could site calibration control points be used for future checks? 8. What is the appropriate number and configuration of control points for site calibration? Can the 2007 specification be reduced? Under the 2007 specification, there were only two project control points that could be used to check site calibrations. 9. Concerning the specified 20 subgrade check points per mile, this was a divided highway. Was the specified number per lane-mile or per linear mile? 9

18 10. Control points are used for other than GPS site calibration (e.g., checking pipe inverts), so dense control was appropriate. 11. Should initial staking and subgrade be separated in the specification? 12. Not all machines have GPS machine guidance, so initial staking is necessary. In addition, subcontractors (e.g., for clearing and grubbing) might not have the technology and need stakes. However, it might be possible to reduce initial staking to every other station (as a test) on the 2008 pilots. 13. The 2007 guidance language was not used on the project because of the experience of both the field engineer and the contractor. However, after the project, the guidance language was reviewed with the full weight of the project experience in mind Project (STH 57) WisDOT Northeast Region Field Engineer: Kevin Derenne Contractor: Mashuda Contractors Contractor Project Manager: Tom Dobberthein Contractor Project Foremen: Matt Mashuda and Mike Burt This project involved improving 4.9 miles of STH 57 from bi-directional to divided highway in southern Door County. The project ran from Truway Road at station to CTH N at station It abutted the Hoffman project which lay to the south. The project had cuts as deep as 18 feet and fills as high as 15 feet. A 100-acre mitigation site was included in the project. The project was not an official pilot because the specification was not incorporated into the contract. However, the contractor tested and used GPS machine guidance on the project and some useful information was obtained. Mashuda has been using RTK GPS for surveys for the past three years. However, this project involved their first use of GPS machine guidance on construction equipment. Mashuda did side-by-side testing, on the project site, of Trimble and TOPCON on two dozers at the same time. Trimble was selected and installed on two dozers. They were used on slopes and ditches only (see Figure 6). The field engineer did not have experience with GPS machine guidance. He kept an electronic diary, including entries pertinent to GPS machine guidance aspects of the job. Weekly project meetings were held in the field on Wednesday afternoons. The project schedule called for staging, such that grading operations were underway for much of the project. Mashuda subcontracted for staking on the project. Checks, against design, were made on slopes and ditches. Check values were consistently much less than ±0.2 ft. The region provided breakline data in.dwg format and Mashuda subcontracted the 3D modeling to POB out of Stevens Point. POB built a single complete model for the entire project. Both Trimble and TOPCON were able to import these data during the side-byside testing mentioned earlier. 10

19 Figure 6. Slope and Ditch Work Using GPS Machine Guidance on the Mashuda Project Discussion Points Raised on the Project 1. Is 3D model building going to be a bid item? Perhaps the initial 3D model, and some number of planned changes, should be a bid item and blue top staking should be eliminated as a bid item. 6. Southwest Region (Madison) Pilot Project 6.1. Pilot Project (STH 106) WisDOT Southwest Region Project Manager: Teri Schopp WisDOT Southwest Region Field Engineer: Jeff Kaarto WisDOT Southwest Region Surveyor: Pat Krebs Contractor: Wondra Excavating Contractor Project Manager: Bob Mayer Contractor Project Foreman: Matt Cameron This project involved improving 9.9 miles of STH 106 in Jefferson County, just east of Ft. Atkinson. The project ran from Edgewater Road at station to CTH CI at station STH 106 is a bi-directional, two-lane roadway. The project consisted of six consecutive segments: 1) 2.5 miles of vertical alignment change and grading; 2) one mile of pavement replacement with no grading; 3) two miles of vertical alignment change and grading; 4) one mile of pavement replacement with no grading; 5) two miles of vertical alignment change with grading; and 6) 1.5 miles of pavement replacement with no grading. There are areas where excavation below subgrade (EBS) was required because of soft soil conditions. Maximum cut and fill depths and heights were nine feet. One mile of urban grading, within Ft. Atkinson, was also included. 11

20 The Southwest Region had no staff familiar with the LANDXML data exchange format required by the 2007 specification. They attempted unsuccessfully to import breakline and mass point data provided by WisDOT Central Office into CAiCE. As a result of the region not being able to examine the data, they served only as a conduit for data exchange between WisDOT Central Office and Wondra. The field engineer and region staff could not view the 3D model because they did not have the necessary software (TOPCON 3D Office). The region project manager and field engineer did not have experience with GPS machine guidance. Therefore, the region elected to modify the specification provided for the 2007 pilot projects. The region s modification affected only subgrade staking aspects of the specification (see Appendix B). The first 3000 feet of the first grading segment required subgrade stakes, with the required number being reduced every 1000 feet. If the finished subgrade met tolerance at subgrade stakes in the first 3000 feet of the first segment, then the engineer could waive requirements for subgrade stakes on the remainder of the project. Wondra agreed to this modification of the specification. The region surveyor developed geodetic control for the project and also placed slope stakes. RTK GPS methods were used for establishing control coordinates and for slope staking. The geodetic control was adopted from six-year-old control and painted photogrammetric targets that had been placed for an earlier mapping flight over the roadway. The existing mapping control had been tied to the HARN and, by three-wire leveling, to an NGS benchmark in Ft. Atkinson. All of the mapping control points were expected to be destroyed by the construction. So, the RTK GPS survey was used to establish four additional control points that would not be disturbed. The calibration for the new RTK GPS survey was referenced to a minimum of two three-dimensional points, four horizontal points, and two vertical points in the mapping control. Ultimately, the field engineer was provided with tie sheets and a listing for 23 mapping control points, seven photogrammetric targets, and the four new control points set by the region surveyor. All control points were within the STH 106 right-of-way. The horizontal geodetic datum is NAD 83 (1991). The vertical geodetic datum is NGVD 29. Wondra found the WisDOTprovided control to be sufficient and did not experience a need to densify it further. The field engineer kept a hardcopy grading diary, including entries pertinent to GPS machine guidance aspects of the project. Project weekly meetings were on Wednesday mornings and included discussions of GPS machine guidance work. Wondra had two TOPCON-equipped dozers on the project site (see Figure 7). They had been using RTK GPS for four years and GPS machine guidance since the fall of Wondra had used GPS machine guidance on county highway projects, but this was their first use of it on a WisDOT project. Grading operations on this pilot project began during the first week of May. 12

21 (a) (b) (c) Figure 7. Two Views of One of Wondra s TOPCON-Equipped Dozer (a,b) and Its Tilt Sensor (c) NOTE: The dual antennae, mounted on a single staff above the blade on the longitudinal axis of the machine, give the blade s coordinates, direction, and pitch angle. The roll angle is measured by the tilt sensor mounted on the blade. The tilt sensor is calibrated using a 4-foot carpenter s level, and the calibration is checked every two weeks. Wondra submitted a GPS work plan (Appendix I) and proceeded accordingly. There was a single base station site for the entire ten-mile project. The base station (see Figure 8) was the top of a permanent post driven into the ground at a convenient central location (not an existing control point). Using a permanent post eliminated the need for leveling and centering the antenna above a ground point each workday morning. The base station s coordinates were established during site calibration. The base station s receiver was removed at the end of each day s work for security purposes. The base station s radio range was approximately seven miles. This could be increased with boosters. In an additional attempt to extend the range on the project site, the base station s broadcast antenna was mounted atop a nearby utility pole. The base station s broadcast frequency is not unique. Ordinary radio signals can sometimes interfere with those of the base station. 13

22 (a) (b) Figure 8. Wondra s Base Station (a) with Broadcast Antenna atop a Utility Pole (b) TOPCON recommends that the geographic extent of work from a base station be no more than ten miles. It is expected that this limitation arises from the single scale factor computed during site calibration and from the assumption of consistent atmospheric conditions at the base station and rovers. Site calibration was done for each of the three grading segments. The calibration file for the first segment was appended for the second segment and, then again, appended for the third. Site calibration was checked at least once per day when the base station was set up (see Figure 9). Additional checks were sometimes made during the lunch hour. The project foreman believed this practice to be adequate because TOPCON receives signals from GLONASS satellites in addition to GPS satellites. Wondra elected to subscribe to TOPCON s optional GLONASS service after realizing a ft improvement in accuracy and, more importantly, elimination of down-time due to weak satellite geometry. The project foreman reported that there are usually no fewer than 12 GLONASS and GPS satellites visible to the receivers. Site calibration files were not provided to the field engineer because he had no way of reading them (they were in TOPCON proprietary format). Site calibration check values for the project appear in Appendix J. The largest absolute value in a horizontal check is 0.04 ft (tolerance = 0.10 ft). The largest absolute value in a vertical check is 0.04 ft (tolerance = 0.05 ft). The project foreman was unable to import the WisDOT-provided LANDXML breaklines and mass points into TOPCON s software. Attempts produced multiple computer errors. The TOPCON vendor was also unable to import the WisDOT breaklines using two different versions of the Carlson CAD software. Horizontal alignments in LANDXML could be imported, but vertical alignments showed only VPIs with zero-length vertical curves. To move forward, the project foreman developed 3D models directly from the 14

23 plans, using Autocad and TOPCON software. Data provided by WisDOT s Central Office was not used. The 3D models were built on-the-fly to stay ahead of the construction crews. Figure 10 provides four views of Wondra s 3D model, two being those of the machine operator. Figure 9. Matt Cameron and Brad Cunningham Making a Site Calibration Check Wondra did not find any problems that required revisions to the project plans. There were minor changes such as saving trees that were designated for removal and moving of residential driveways, but none of these required revisions to the plans. Also, there were some minor design changes that were not incorporated into the 3D model. For example, an existing 80-ft culvert was encountered that did not appear in the design. To include the culvert in the 3D model, it would have been necessary to conduct a survey and reduce the data for inclusion in the model. It was easier to merely ask the machine operator to make the grade meet the culvert. In the first 1000 ft of blue-topping, three stakes were set at each full station. In the second 1000 ft, every full station on only the centerline was staked. At the first two full stations, the subgrade was consistently 0.2 ft below the blue tops. No blade offset had been used during grading in this area because dirt was being removed (cut). Blade offsets are often used in fill areas that are expected to compact. The project foreman believes the material was so soft in this area that compaction occurred and a blade offset should have been used (see Figure 11). Appropriate blade offsets vary with soil conditions. The Hoffman foreman reported that blade offsets from 0.02 to 0.10 ft. are often used. Because the blue tops were missed at the first two stations, the project engineer decided to not completely waive blue-topping for the remainder of the project. On straightaways the remaining centerline was staked every 500 ft. All superelevated curves were staked three-across at all full stations. 15

24 Figure 10. Four Views of the Wondra 3D Model (Bottom Photos are inside the Cab of a Dozer) Appendix K contains the full set of 114 subgrade checks for the project. The largest positive subgrade check difference is ft. The largest negative subgrade check difference is ft. The mean and standard deviation of the subgrade checks are ft and ±0.05 ft, respectively. In only a single instance on the Wondra project did more than one of any five consecutive subgrade checks exceed 0.10 ft in absolute value. The two offending checks were separated by 1000 feet horizontally. Sta (12.0L) checked at +.11 ft and Sta (12.0L) checked at ft. Between these two checks, there were three checks less than 0.10 ft. Figure 12 shows the finished subgrade and slopes with and without base course placement on the Wondra project Discussion Points Raised on the Project 1. Data management issues were prevalent and need to be addressed. 2. Should blue-topping be at the option of the field engineer? 3. Is it necessary to check the site calibration twice a day? 16

25 Figure 11. Soft Material on the Wondra Project Site (a) (b) Figure 12. Finished Subgrade without (a) and with (b) Base Course Placement On the Wondra Project 4. What are the appropriate ways for the contractor and the department to share electronic information in proprietary formats (e.g., site calibration files). 5. Last-pass data collected by the machines is not an efficient method for collecting as-built data. Some last-passes have to be passed over again because of rutting. Some slopes have smaller footprints than the machines. Collected data are unlabelled points. Breaklines would have to be constructed later. It would be difficult to track what was done, what needed to be done and what needed to be done over. 6. Slope staking should not be eliminated or even reduced. For example, topsoil stripping on this project was done without GPS machine guidance. The terrain was irregular with many transitions between cuts and fills. Therefore, the slope stakes were not uniform. Eliminating every other slope stake could lead to under-stripping or over-stripping. 17

26 7. Summary of Issues for Discussion 1. What will be WisDOT s commitment on provision of design-side data? Considerations include limited resources, the period of transition from CAiCE to Civil 3D, and allotted time for making changes (e.g., some changes require more time than others). There is an urgent need to address these and other data management issues such as whether or not 3D modeling should be a bid item, the impact of involvement of third parties (subcontractors for 3D modeling), and the role of the field engineer when not equipped with necessary software. 2. What is the best means for getting other WisDOT regions involved? 3. What effective educational mechanisms are there for WisDOT and contractor staff? 4. Will future GPS machine guidance grading specifications include slopes and ditches or will they continue to be confined to roadway subgrade? Are there aspects of the standard specifications that affect grading of slopes and ditches and should be addressed for GPS machine guidance? 5. When WisDOT goes to statewide options for bidding, will the regions have choices or will the option apply to all contracts? 6. Is it possible to use existing section corners for horizontal control? 7. There are problems with using last-pass data from the machines to construct as-built surfaces. For example, data are collected whether the blade is cutting, flush with the ground, or above ground. Also, such data can be collected only as points. No breaklines or labels can be inserted, so the data would have to be interpreted and edited later with no information other than visual inspection on which points should be connected to form lines. Furthermore, some last-pass data have to be passed over again (for example, because of rutting). 8. The 2007 specification allows the engineer to require conventional staking (blue tops) if GPS machine guidance is producing unacceptable results. Should there be more flexibility for the engineer (e.g., requiring some blue topping by choice)? Should the contractor also have the option of going to conventional staking (e.g., equipment priorities)? 9. Are 2007 tolerances too stringent or not stringent enough? Feedback from the projects indicated they are appropriate and should not be changed. 10. Why is it necessary for the site calibration check points to be different from the control points used for calibration? If the site calibration checks once, could site calibration control points be used for future checks? Can site calibration checks be reduced to one per day? 18

27 11. What is the appropriate number and configuration of control points for site calibration? Can the 2007 specification be reduced? Under the 2007 specification on the Hoffman project, there were only two project control points that could be used to check site calibrations. 12. Concerning the specified 20 subgrade check points per mile, what about divided highways? Is the specified number per lane-mile or per linear mile? 13. Should initial staking and subgrade be separated in the specification? 14. Not all machines have GPS machine guidance, so initial staking is necessary. In addition, subcontractors (e.g., for clearing and grubbing) might not have the technology and will need stakes. There are mixed opinions on whether or not slope staking can be reduced to every other station. Doing so could cause problems with, for example, topsoil stripping in non-uniform areas. 15. What are the appropriate ways for the contractor and the department to share electronic information in proprietary formats (e.g., site calibration files). This is separate from the 3D modeling data management issue. 8. Lessons Learned from 2007 Pilots Shortly after submittal of this project s interim report (Vonderohe (2007b)), the project s Advisory Group was convened to consider issues raised during the pilot projects and make recommendations concerning revisions to the specification and guidance language. Those in attendance when the Advisory Group met were Ken Brockman (WisDOT), Jerry Zogg (WisDOT), Alan Rommel (WisDOT), Rick Larson (WisDOT), Mike Hall (WisDOT), Brad Hollister (WisDOT), Kris Sommers (WisDOT), Chris Goss (Hoffman Construction), Mike Bradley (Ayres & Associates), Matt Cameron (Wondra Excavating), and the author. Because of the pervasiveness and depth of issues surrounding development and management of 3D models, a second Data Management Group was convened. Those in attendance at the time of the meeting were Ken Brockman (WisDOT), Jerry Zogg (WisDOT), Alan Rommel (WisDOT), Rick Larson (WisDOT), Brad Hollister (WisDOT), Kris Sommers (WisDOT), Chris Goss (Hoffman Construction), Matt Cameron (Wondra Excavating), Jason Pingel (POB), Mike Larenzo (Kapur & Associates), and the author. As these groups discussed the issues, it became apparent that 1. The specified accuracy requirements can be met with GPS machine guidance. 2. There is still a need to do slope staking as slope stakes are used for visual reference and for applications other than grading. The question remains whether or not the frequency of slope staking can be reduced. 3. There is no need to set blue tops, but the department should continue to make independent checks of the subgrade. 19

28 4. The specification requires the contractor to make a GPS rover available to the engineer and to provide training. There is a need for a better longer-term solution for the engineer s access to the technology. 5. There is a need to clarify some aspects of the specification (e.g., minimum frequency of subgrade checks for divided highways. 6. Formats for data exchange should be made more flexible. 7. The capacity of WisDOT for provision of preliminary data for 3D modeling, through the Methods Development Unit, will be about 15 projects per year as the transition is being made to 3D design. 8. Roles and responsibilities, concerning 3D model data development and management, of central office staff, region staff, and contractors need to be well-described and understood. 9. The impact of WisDOT s continuously-operating reference stations (CORS) on GPS machine guidance has yet to be studied. 10. There are impediments to using last pass data from the machines to form as-built surface models for determining final quantities. These impediments need to be better understood. 9. Revisions to Specification Based upon the deliberations and recommendations of the Advisory and Data Management Groups, some revisions were made to the 2007 specification. These include: 1. Reducing the required frequency of site calibration checks from one every five operating hours to one per day. 2. Adding language to emphasize that subgrade checks are to be made against plan elevations (i.e., elevations shown on the paper plans). 3. Specifying that a minimum of 20 subgrade checks per roadway mile are required. 4. Including.dwg or other engineer-approved format as acceptable for 3D data exchange. 5. Because the 2008 projects are bid with the specification for GPS machine guidance, the contractor is provided with the option of using conventional equipment, GPS machine guidance, or a combination of the two. The 2008 pilot specification appears in Appendix L. 20

29 10. Revisions to Guidance Language Based upon the deliberations and recommendations of the Advisory and Data Management Groups, some revisions were made to the 2007 guidance language. These include: 1. Describing the roles and responsibilities of central office staff and region staff in development and management of 3D model data. 2. Adding language concerning the engineer s responsibility in developing an archive of 3D model revisions. 3. Stating that the engineer should use technologies independent of the contractor s GPS installation to make initial subgrade checks. Total stations and / or differential leveling should be used until the engineer is satisfied that the GPS machine guidance technology is meeting requirements. The 2008 pilot guidance language appears in Appendix M. 11. Summary of Issues Intended to be Addressed by the 2007 Pilot Projects Here, the general questions identified in Section 3 are repeated along with notes on their status as a result of the 2007 pilot projects: Equipment: 1. What are the frequency, duration, and types of problems with operation of the technology (e.g., poor satellite geometry, loss of lock, multipath, software glitches, data entry and other human errors, technology incompatibilities)? Systems that receive only GPS signals sometimes experience nearlyinconsequential downtimes of minutes per day. Systems that receive GLONASS signals in addition to GPS signals experience less downtime. Multipath and GPS software glitches do not appear to cause problems. Final pass blade depths should be set to account for compaction. Unless resolved, data format incompatibilities inhibit timely data exchange. 2. What are the vertical tolerances that are achievable using GPS Machine guidance? The pilot specification tolerance of no more than one out of any five consecutive check points not exceeding 0.1 ft appears to be appropriate. Further monitoring on the 2008 pilot projects is recommended. 3. What are the necessary knowledge and skill levels for project engineers, contractor project managers, and machine operators? Machine operators need to be skilled in use of the technology, but probably do not need a great deal of understanding in how the technology works. Project managers and project engineers should not only know how to use the technology but also have enough background in its workings to judge data quality and solve problems that might arise. This issue needs further exploration to develop detailed recommendations. 4. What other efficiencies are realized with GPS machine guidance? 21

30 Less repeat work, less human judgment (subject to error), more visuallypleasing results. 5. What other difficulties arose with GPS machine guidance? No severe difficulties, beyond data exchange, were encountered. Further monitoring on the 2008 pilots is recommended. 6. Is it necessary for the height modernization program to be completed in the area for GPS machine guidance grading to be successful? If not, what additional project control is needed and what other challenges are encountered? None of the control for either project was tied to the height modernization network. The Wondra project had most of its control already in place from previous WisDOT work. The Hoffman project had its control tied to previously-established control points other than height modernization. Department Responsibilities: 1. Are 3D model components provided by WisDOT readily usable by contractors? They were not until the parties began using.dwg format. LANDXML is not an industry-wide uniform standard. 2. What are the frequency and causes of revisions to three-dimensional models? Very few, if any, revisions were made to the models after initial development by the contractors and checking by WisDOT. No revisions to project plans were necessary on either pilot project. Further monitoring on the 2008 pilots is recommended. 3. Are data exchange standards and rates sufficient for updating models during construction? Models did not need updating because there were no plan revisions. Further monitoring on the 2008 pilots is recommended. 4. What is the appropriate spatial frequency for quality assurance checks and what are the appropriate tolerances? Checks made by the project engineers were not monitored. This question needs to be investigated thoroughly on the 2008 pilot projects. 5. Are there issues, during construction, with utility coordination, subcontractors, or others due to reduced staking requirements? Both pilot projects were fully slope staked. Contractor Responsibilities: 1. What is the appropriate control configuration for GPS site calibration? The pilot specification of six control points or two per mile, whichever is greater, worked. However, further monitoring on the 2008 pilots is recommended. 2. What are the appropriate tolerances for GPS site calibration? The site calibration check tolerances of 0.1 ft horizontally and 0.05 ft vertically appear to be appropriate. Further monitoring on the 2008 pilot projects is recommended. 3. What is the appropriate frequency for GPS site calibration checking / recalibration? It appears that one site calibration check per day is adequate. 22

31 4. What is the maximum geographic extent over which a single GPS site calibration is valid? This question was not adequately addressed on the 2007 pilots. It is recommended that further investigation be done during What is the appropriate spatial frequency for quality control checks and what are the appropriate tolerances? The pilot specification tolerance of 20 or more checkpoints per roadway mile and no more than one out of any five consecutive check points not exceeding 0.1 ft appears to be appropriate. Further monitoring on the 2008 pilot projects is recommended. 6. What needs to be staked and what staking can be eliminated? It is apparent that blue top staking can be eliminated. Slope staking is necessary for visual reference and for applications other grading. However, under some circumstances, it might be possible to reduce the frequency of slope staking. 7. What is the appropriate number of supplemental control points to be furnished by the contractor and what are the appropriate methods for development of supplemental control? Neither contractor needed to develop supplemental control. The control provided by WisDOT was adequate. 8. Do minor field changes need to be incorporated in the 3D model or is it more cost effective to have the machine operator override automatic controls for minor changes? The contractors found it more cost effective to make minor revisions in the field only, without revising the 3D models. 9. What types of field changes warrant changes in the 3D model? Although not fully addressed, it appears that only field changes leading to revisions in the plans warrant changes in the 3D model. 12. Plans for Potential 2008 Pilot Projects As of the date of this report, the regions have proposed the following six projects as potential pilots for 2008: NC Region: 1. Project ID Kowalski Rd Overpass length = 0.5 miles Concept = Overpass expansion project (2 lanes to 4 lanes) rural grading. Let = 3/11/08 2. Project ID /76 USH 51/STH 29 Corridor- Wausau (Marathon County) Length = 0.9 miles urban (28 th Ave and Stewart Ave), 0.9 miles total rural (all ramps) Concept = Reconstruction project, structures, rural and urban grading. Let = 2/1/08 23

32 SE Region (Note: This project is within the CORS Phase 1 area): 3. Project ID Burlington Bypass - STH 36 (South) to STH 83 (South) (Racine County) Length = 2.7 miles Concept = Grading Let = 11/13/07 (H James and Sons is contractor, Region is discussing a no cost CCO). NE Region (Note: These three projects are within the CORS Phase 1 area) 4. Project ID USH 41, Oconto Bypass (Oconto County) Length = 1.75 miles Concept = Grading Let = 3/11/08 5. Project ID USH 41, Oconto Bypass (Oconto County) Length = 3.11 miles Concept = grading Let = 4/08/08 6. Project ID USH 41 Pestigo Bypass (Marinette County) Length = 3.74 miles Concept = Grading Let = 2/12/08 Of these, one has already been bid and would be managed by change order. The other five have yet to be let and, consequently, would be bid under the 2008 specification at the option of the bidder. It is anticipated that more potential 2008 pilots will be identified and that not necessarily all identified projects will be bid for GPS machine guidance. Ultimately, the desirable number of 2008 pilot projects is Issues to be Addressed during 2008 Pilot Projects Among the issues to be addressed during the 2008 pilot projects are: 1. What is the impact of WisDOT s CORS network on GPS machine guidance? Five of the projects identified above are within the CORS Phase 1 area. Phase 1 is expected to become operational during April, 2008, in time to be used during the construction season. Pilot project contractors would need to adopt CORS technology and services if their projects were to be used to address this issue. 2. Are the described roles and responsibilities for development, exchange, and revisions to 3D model data appropriate? 3. Are there other workflow, resource, or technological impediments to development and management of 3D model data? 24

33 4. Does the specification continue to perform adequately in the areas of GPS work plan, site calibration, site calibration checking, and subgrade checking? 5. Can project control requirements be reduced? 6. Is there a need to specify the maximum geographic extent of a single site calibration? 7. Can frequency of slope staking be reduced? Is the answer to this question site-dependent? 8. Can the amount of information placed on slope stakes be reduced? 9. What are the options for longer-term solutions to the GPS-rover-access problem for the engineer? 10. Are there training or educational needs that are not being addressed? 11. What are the details of the impediments to using last pass data from the machines for as-built purposes? The 2008 specification is silent on last pass data, but if contractors are interested in pursuing this question, then it bears consideration. 12. Are there issues with GPS machine guidance in urban areas that are not apparent in rural areas? 13. What can be learned from the experience of having GPS machine guidance as a bid option? 14. What is the appropriate spatial frequency for additional quality assurance checks by the project engineer and what are the appropriate tolerances? Outreach During 2007, the project team gave presentations on background, status, and plans for implementation of GPS machine guidance at 1) the open house sponsored by the Northeast Region, 2) two meetings of the WisDOT / WEMA Grading, Landscaping, and Sewer Committee, 3) a monthly WisDOT Project Development Staff meeting, and 4) the annual meeting of the Wisconsin Earth Movers Association. An additional presentation is planned for the 2008 annual meeting of WisDOT / WTBA in January. The 2007 final report (Vonderohe (2007a)) is available through the Construction Materials and Support Center s web site ( This report is also expected to be available through that website. 13. The Bigger Picture A Period of Transition Technology is constantly changing while institutions and cultures attempt to keep up. Implementation of GPS machine guidance for highway construction is engaged in a rapidly-evolving period of transition. Changes are being made on three major fronts: 25

34 1. The adoption of GPS machine guidance itself is a new way of doing business for both industry and government, with semi-intelligent machines knowing where they are in both real and virtual worlds. 2. The Global Positioning System itself is now being described as part of a larger Global Navigation Satellite System (GNSS) that includes the Russian GLONASS and European Galileo systems. All components of GNSS are undergoing technological upgrades. In addition, the revolutionary concept and technology of CORS for RTK GPS positioning have emerged in the last two years. 3. The engineering community, including WisDOT, is migrating from twodimensional to three-dimensional design technology and thinking. This transition requires not only a change in the way of doing business but also a change in way we think about and describe the world. These are exciting times. They are also challenging times. Human, institutional, and cultural changes are not easy. They take time. Technological change is easy. It happens over night. Whether or not we can more fully realize the true benefits of technological advances depends upon our ability to adapt our institutions, our culture, and our individual and collective thinking. Acknowledgements The author gratefully acknowledges the support of WisDOT and the CMSC at UW- Madison. Many project personnel, both contractor employees and departmental staff, contributed valuable assistance and information for this report. The Advisory and Data Management Groups provided guidance and reached consensus on revisions to the specification and guidance language. Project objectives could not have been met without their dedicated efforts. Rick Larson, Mike Hall, Kris Sommers, and Alan Rommel developed much of the detailed language in the revised (2008) specification and guidance language. Trimble and TOPCON provided software. List of References Vonderohe, A.P., (2007a), Implementation of GPS Controlled Highway Construction Equipment, Final Report submitted to the Wisconsin Department of Transportation, Construction and Materials Support Center, University of Wisconsin Madison, April. Vonderohe, A.P., (2007b), Implementation of GPS Controlled Highway Construction Equipment Phase II, Interim Report submitted to the Wisconsin Department of Transportation, Construction and Materials Support Center, University of Wisconsin Madison, August. 26

35 Appendix A. Specification and Guidance Language for 2007 Pilot Projects (As Presented in Vonderohe (2007)) 27

36 Specification and Guidance for 2007 Wisconsin Department of Transportation GPS Machine Guidance Pilot Projects NOTE: This document is intended for use on only the 2007 pilot projects for GPS machine guidance. The specification and guidance language are subject to change for the 2008 and later construction seasons. In addition, this document is subject to possible modification by WisDOT regions with 2007 pilot projects PILOT PROJECT SPECIFICATION Construction Staking Subgrade, Item ; Construction Staking Initial Layout, Item Conform to standard spec 650 as modified in this special provision. Replace standard spec with the following: Subgrade General (1) Use global positioning system (GPS) machine guidance or conventional subgrade staking on designated portions of the contract as follows: GPS Machine Guidance Subgrade Staking (2) The engineer may require the contractor to revert to conventional subgrade staking methods for all or part of the work at any point during construction if, in the engineer's opinion, the GPS machine guidance is producing unacceptable results. If the engineer revokes approval to use GPS machine guidance on all or part of the work for reasons beyond the contractor's control, the department will measure the additional subgrade staking required to successfully complete the work in those areas as specified in 650.4(2) of this special provision Subgrade Staking (1) Set construction stakes or marks at intervals of 100 feet, or more frequently, for rural sections and at intervals of 50 feet, or more frequently, for urban sections. Include additional stakes at each cross-section as necessary to match the plan cross-section, achieve the required accuracy, and to support construction operations. Also set and maintain stakes as necessary to establish the horizontal and vertical positions of intersecting road radii, auxiliary lanes, horizontal and vertical curves, and curve transitions. Locate stakes to within 28

37 0.25 feet (75 mm) horizontally and establish the grade elevation to within 0.03 feet (10 mm) vertically GPS Machine Guidance General (1) No subgrade stakes are required for work approved for GPS machine guidance. (2) Coordinate with the engineer throughout the course of construction to ensure that work performed using GPS machine guidance conforms to the contract tolerances and that the methods employed conform to the contractor's GPS work plan and accepted industry standards. Address GPS machine guidance issues at weekly progress meetings. (3) Provide GPS rover equipment to department staff as requested to check the work. This equipment is not intended for exclusive use of the department and may be used by the contractor as needed on the project. Provide training for department staff designated to use the GPS rover. Training shall include but not be limited to hardware, software, and operation of GPS rover equipment. Provide a copy of the user manual for the supplied rover equipment. Provide routine maintenance of equipment provided for department use. The department is responsible for loss of, or damage (beyond normal wear and tear) to, the rover while in the engineer's possession GPS Work Plan (1) Submit a comprehensive written GPS work plan for department review at least 10 business days before affected grading operations begin. The engineer will review the plan to determine if it conforms to the requirements of this special provision. (2) Construct the subgrade as the contractor's GPS work plan provides. Update the plan as necessary during construction of the subgrade. (3) The GPS work plan should discuss how GPS machine guidance technology will be integrated into other technologies employed on the project. Include, but do not limit the contents to, the following: 1. Describe the manufacturer, model, and software version of the GPS equipment. 2. Provide information on the qualifications of contractor staff. Include formal training and field experience. Designate a single staff person as the primary contact for GPS technology issues. 3. Describe how project control is to be established. Include a list and map showing control points enveloping the site. 4. Describe site calibration procedures. Include a map of the control points used for site calibration and control points used to check the site calibration. Describe the site calibration and checking frequency as 29

38 well as how the site calibration and checking information are to be documented. 5. Describe the contractor's quality control procedures. Describe procedures for checking, mechanical calibration, and maintenance of equipment. Include the frequency and type of checks performed to ensure that the constructed subgrade conforms to the contract plans Equipment (1) Use GPS machine guidance equipment to meet the requirements of the contract. (2) Perform periodic sensor calibrations, checks for blade wear, and other routine adjustments as required to ensure that the final subgrade conforms to the contract plans Geometric and Surface Information Department Responsibilities (1) The department will provide to the contractor the best available electronic files of survey and design information as described here in and in CMM The department incurs no additional liability, beyond that specified in standard spec or standard spec 650, by having provided this additional information. (2) The department will provide data to the contractor that include the following: Data Type Reference Line Data Design Profile Data Proposed Cross Section Data Existing Surface DTM Data Existing Topographic Data (excluding utilities) Superelevation Data Graphical Information Format LandXML LandXML Land XML LandXML DTM LandXML LandXML DGN or DWG (3) The department will provide design surface data in the form of points and break lines derived from the cross sections in the contract in LandXML format. The points and break lines will be on the subgrade surface between the subgrade shoulder points, and will be on the finished surface in topsoiled areas. The department provides design surface data for information only, and has no contractual liability for it. 30

39 Contractor Responsibilities (1) Develop and maintain the initial design surface DTM for areas of the project employing GPS machine guidance consistent with information the department provides. Confirm that the design surface DTM agrees with the contract plans. (2) Provide design surface DTM information to the department in LandXML or other engineer-approved format Managing and Updating Information (1) The department and contractor will agree on the design surface model before using it for construction. Provide a copy of the resultant design surface DTM to the engineer at least two business days before using that design surface DTM for construction. (2) Notify the department of any errors or discrepancies in department-provided information. Provide information regarding errors or discrepancies in the existing surface DTM, identified in the field, to the department in LandXML format if a revision to the contract plans is required. If surveying work beyond that required to slope stake is required to re-define the existing surface, the department will pay for costs of that additional surveying as extra work. (3 The department will determine what revisions may be required. The department will revise the contract plans and existing surface DTM, if necessary, to address errors or discrepancies that the contractor identifies. The department will provide the best available electronic files and other available information related to those contract plan revisions. (4) Revise the design surface DTM as required to support construction operations and to reflect any contract plan revisions the department makes. Perform checks to confirm that the revised design surface DTM agrees with the contract plan revisions. Provide a copy of the resultant revised design surface DTM to the engineer. The department will pay for costs incurred to incorporate contract plan revisions as extra work. (5) The department will maintain the existing surface DTM by incorporating needed revisions. The department will make the current existing surface DTM available, in LandXML DTM format, to the contractor throughout construction Site Calibration (1) Designate a set of control points, including a total of at least 6 horizontal and vertical points or 2 per mile, whichever is greater, for site calibration for the portion of the project employing GPS machine guidance. Incorporate the department-provided control framework used for the original survey and design. 31

40 (2) Calibrate the site by determining the parameters governing the transformation of GPS information into the project coordinate system. Provide the resulting site calibration file to the engineer before beginning subgrade construction operations. (3) In addition to the site calibration, perform site calibration checks. Perform these checks at individual control points not used in the initial site calibration. At a minimum, check the calibration at the start of each day and at least once for every 5 hours of continuous subgrade construction work. Report out-oftolerance checks to the engineer. The measured position must match the established position at each individual control point within the following tolerances: - Horizontally to 0.10 feet or less. - Vertically to 0.05 feet or less. (4) Provide the previous week s daily calibration check results to the engineer at the weekly progress meeting for monitoring the GPS work. (5) The department will use the same calibration file the contractor uses Construction Checks (1) Conduct calibration checks daily conforming to of this special provision and consistent with the contractor's GPS work plan. Use a GPS rover to check the subgrade at 20 or more randomly selected locations per mile. Document all GPS rover subgrade checks and any auxiliary checks made using other technologies. Provide all documentation to the engineer. (2) Ensure that at least 4 of any 5 consecutively-tested subgrade points are within 0.10 foot vertically of the plan elevation. Notify the engineer if more than one of any five consecutively-tested subgrade points differs by more than 0.10 feet from the plan elevation. (3) The department will conduct periodic independent subgrade checks using the contractor supplied GPS rover or conventional survey methods. When using the GPS rover, the department will use the same calibration files and other hardware and software settings the contractor uses for subgrade checking. The department will notify the contractor if any individual check differs by more than 0.10 feet from the design. Replace standard spec with the following: Initial Layout (1) Set and maintain construction marks as required to support the method of operations consistent with third-order, class I horizontal and third-order vertical accuracy. Validate the department-provided horizontal and vertical control information and notify the engineer of any discrepancies. Provide 32

41 marks to establish and maintain intermediate vertical and horizontal control for reference line alignment, side road alignments, radius points, bench level circuits, slopes on the ground, and offsetting the horizontal roadway alignment. These marks constitute the field control used to govern and execute the work. (2) For the portion of the project using GPS machine guidance, set and maintain supplemental control points sufficient to ensure that there are a minimum of 6 established control points per mile. Ensure that these control points are consistent with third-order, class I horizontal and third-order vertical accuracy. Establish vertical control by differential leveling. (3) Verify the existing ground elevations as shown for all roadways on crosssection sheets for accuracy. If the elevation at the slope intercept is off by more than 0.4 foot (120 mm), notify the engineer. Take and document a minimum of 3 shots per roadway section. Set and maintain slope stakes on each side of the road at each cross-section location shown on the plans. Stake additional clearing and grubbing, and marsh excavation limits at locations where they vary from the slope stakes. (4) Document and provide to the engineer complete descriptions and reference ties for the control points, alignment points, and benchmarks to allow for quick reestablishment of the plan data at any time during construction and upon project completion. Replace standard spec with the following: Measurement (1) The department will measure the Construction Staking bid items for base, concrete pavement, resurfacing reference, and initial layout by the linear foot acceptably completed, measured along each roadway centerline. The department will not measure construction staking for base underlying concrete pavement. (2) The department will measure Construction Staking Subgrade by the linear foot of subgrade acceptably completed, measured along each roadway centerline. The department will base measurement on the length of acceptably completed subgrade whether that subgrade was constructed using GPS machine guidance or using conventional construction staking. The department will include the length of subgrade where GPS machine guidance is initially employed but subsequently suspended by the engineer for reasons beyond the contractor's control. The department will measure this work twice, once for the suspended GPS work and once for the conventional subgrade staking required to successfully complete the work. If the department suspends GPS work for reasons within the contractor's control, the department will measure work in the affected area only once. 33

42 (3) The department will measure Construction Staking Curb Gutter and Curb & Gutter by the linear foot acceptably completed, measured along the base of the curb face. The department will measure Construction Staking Concrete Barrier by the linear foot acceptably completed, measured along the base of the barrier. The department will not measure these bid items if abutting concrete pavement. (4) The department will measure Construction Staking Storm Sewer System as each individual inlet catch basin, manhole, and endwall acceptably completed. (5) The department will measure Construction Staking Pipe Culverts by each individual pipe culvert staked and acceptably completed. (6) The department will measure Construction Staking Structure Layout as a single lump sum unit for each structure acceptably completed. The department will measure Construction Staking Electrical Installations as a single lump sum unit for all electrical installations acceptably completed on each project. 34

43 GUIDANCE FOR 2007 WISCONSIN DEPARTMENT OF TRANSPORTATION GPS MACHINE GUIDANCE PILOT PROJECTS GENERAL CONSTRUCTION AND PROJECT SELECTION The candidate project should first be reviewed for suitability for GPS use; for example, projects with dense tree canopy or large vertical cuts may not prove suitable. The region surveyor would assist in this preliminary evaluation with the construction engineer. It may also be determined that only certain project segments would be suitable. Recommended pilot projects should be communicated to the region s project manager and forwarded to Ken Brockman, Bureau of Project Development (BPD) for final approval. On the pilot projects, the item of machine control grading will be used to replace subgrade staking on the whole project or segments of selected roadway sections. The project or segments should be reviewed and agreed upon by the engineer and contractor. A no-cost change order would then be submitted to allow the machine control grading. The item for Staking Subgrade would be paid for in all segments where machine control is attempted. It is recommended that projects using machine control grading would also include contractor staking items such as initial layout. DESCRIBING PROJECT EXTENTS The GPS machine guidance pilot project specification allows some or all of the construction project to be done with GPS machine guidance. If the entire project is to be done with GPS machine guidance, then the following location description table can be used: GPS Machine Guidance Subgrade Staking Entire Project None If segments of the project are to be done with GPS machine guidance and the remaining segments are to be done using conventional construction methods, the segments using conventional methods must be subgrade staked. The extents of each GPS machine guidance segment and each subgrade staking segment need to be described. There are a number of methods for describing the extents of segments. Examples include project stationing (preferred), cross street (intersection) naming, and bridge identification. 35

44 The following location description table combines some of these methods to describe the extents of four segments: GPS Machine Guidance Subgrade Staking From Sta to the From Sta 0+00 to Sta intersection with CTH N. From the intersection with CTH N to the Elm Street overpass (B ) From the Elm Street overpass (B-05-51) to EOJ ROLES AND RESPONSIBILITIES Designer The project designer is responsible for overall design and any subsequent changes. The designer provides normal digital data exchange data including DTM information and would work with the Methods Development Unit (MDU) engineer to prepare XML format information to be used by the contractor. Some additional field verification of models and digital terrain models (DTMs) may be required as a quality assurance of this information. The designer would make the necessary design changes in case of errors and work with the MDU engineer to provide changed DTMs. Construction Engineer Project Selection For the pilot projects, the construction engineer would assist in the determination of the applicability of the use of machine control. The engineer should work with the region surveyor to evaluate the suitability of GPS technology and the availability of project control for the proposed project. The engineer, contractor, and region surveyor should agree on usage and limits of machine control grading, and a recommendation should go to regional & BPD management as noted above. The engineer would lead the coordination of department-provided items and be the focal point for communication with the contractor. Data and Surface Model Coordination In order to prepare project data, DTMs, and surface model information for use by the contractor, there needs to be close coordination between the construction engineer, the designer, and the methods development unit (MDU) engineer. A meeting as noted below could help facilitate this. Initial Coordination Meeting Integral DOT/consultant staff that will provide information and guidance to the project should meet to discuss roles and responsibilities. This should include the design engineer, construction engineer, regional surveyor, methods development engineer, appropriate management, and may include contractor survey personnel. Some of the items to be addressed include provision of models and their formats, survey data and support, and project communications. 36

45 Pre-Construction Survey Meeting Before the start of construction survey, it is recommended that a coordination meeting be held to aid in the passing of survey information to the contractor and discuss the contractors GPS work plan. Pre-Survey Meeting This meeting includes the contractor, contract surveyor, construction engineer, methods development engineer, and regional surveyor. At this meeting, the contractor should share and discuss their GPS work plan, project schedule, and survey schedule. The department should identify key personnel and methods for handling changes on the model, etc. During Construction Calibration checks are the responsibility of the contractor, but should be reviewed with the region surveyor to verify they are in reasonable tolerances and format. The engineer should work with the region surveyor to develop a plan to perform construction checks. It is essential to provide some checks at project start-up to ensure contractor methods are meeting necessary tolerances. These checks can be performed using contractor-supplied GPS rover, independent GPS equipment, or conventional survey methods, and should meet specified tolerances. It is anticipated that once initial methods are working, construction checks could be performed using contractor-supplied rover. The department reserves the right to do added checks as needed. After Construction The contractor, construction engineer, and surveyor should meet to review the effectiveness of the machine control grading operations and identify benefits and issues to be addressed. The construction engineer should prepare a final report evaluating the machine control usage. Evaluation items could include overall project impacts, specification improvements, construction administration issues and other pertinent items. This evaluation should be submitted to the machine control grading steering team; Ken Brockman in the Bureau of Project Development is the designated lead for submittals. Region Surveyor The region surveyor is responsible for providing control points and technical support on the project. Control Points For the pilot projects, the region s survey unit would provide a minimum of 6 control points or 2 points per mile for use during the project. These points should be constructed or located outside the anticipated construction footprint. They should be type 1 or equivalent and should be set 15 degrees clear to the horizon with 360-degree access desirable at 6 foot height. Control points should have horizontal and vertical project coordinates published. These points should be available two weeks before the preconstruction conference. Technical Support The region surveyor should assist in initial evaluation of the project for potential GPS use. The surveyor could also assist in the development of a plan for providing construction checks. The contractor is required to do their own project calibrations and check their work as it progresses. However, there may be questions that arise from the construction engineer 37

46 related to GPS operations and calibrations. It is expected that the regional surveyor would be available to lend technical guidance as warranted. The surveyor should assist in evaluation of the pilot and provision of specific feedback on issues to be resolved, etc. EARTHWORK QUANTITIES The region surveyor should work with the construction engineer and contractor to obtain as-built data and/or construction surface models for computing final quantities. The surveyor would work with the MDU engineer and the construction engineer to develop informational quantities for comparison to conventional quantity computations. SITE CALIBRATION AND CHECKS The contractor performs site calibration and site calibration checks. The contractor provides data collected during these activities to the construction engineer. The following is intended for both the contractor and the construction engineer as guidance in configuring the control points used for site calibration, interpretation of site calibration and check data, and appropriate procedures to follow if either of the specified site calibration check tolerances is exceeded. The construction engineer can also consult with the regional surveyor on these matters. Site Calibration Site calibration, sometimes referred to as localization, for GPS machine guidance is a process that results in computation of parameters for transforming measured GPS coordinates into the coordinate system of the project control points. Good site calibration and checking are vital to the success of GPS machine control operations. Control Point Configuration The GPS machine guidance pilot project specification requires that a minimum of six control points be used for site calibration and that the site calibration be periodically checked at control points not used in the calibration itself. The control points used for site calibration should envelop the project and be well distributed around its perimeter. Control points in close proximity to one another should be avoided. Long, narrow configurations of control points should be avoided. There should be control points near the corners of the project and approximately midway along its boundaries. Error Estimates Horizontal and vertical tolerances are specified for the site calibration checks but not for the site calibration itself. Once the site calibration measurement process is complete, the RTK GPS software will report estimates for the horizontal and vertical errors at each of the site calibration control points. A majority of the horizontal error estimates should be 0.10 feet or less in magnitude. A majority of the vertical error estimates should be 0.05 feet or less in magnitude. If any horizontal error estimate is greater than 0.15 feet, or if any vertical error estimate is greater than 0.08 feet, it is indicative (but not conclusive) that there might be later difficulties in meeting the site calibration check tolerances at the independent control points. These tolerances are 0.10 feet (horizontal) and 0.05 (vertical). Site Calibration Checks If any site calibration check exceeds the specified tolerances (i.e., 0.10 horizontally or 0.05 feet vertically), there is a sequence of steps that should be followed: 1. The check should be re-measured at the same independent control point to ensure there is no problem with the check measurement. 38

47 2. A second and, perhaps, a third independent control point should be used to check the site calibration. If tolerances are met at these additional independent control points, then a problem is indicated with the first check control point. 3. If check tolerances are not met at two or more independent control points, then a problem is indicated with the site calibration and the site calibration measurement and computation procedure should be repeated to ensure that there is no problem with the initial site calibration measurements. If site calibration problems persist, vendor supplied manuals or guidance might also need to be consulted. 4. If the repeated site calibration measurements are in close agreement with the initial site calibration measurements, then a problem is indicated with one or more of the site calibration control points. The site calibration should then be performed while excluding the control point with the largest horizontal and / or vertical error estimate. It is likely that such error estimates will be larger than 0.10 feet horizontally or 0.05 feet vertically. 5. If a problem with a site calibration control point is identified in step 4, that control point should be replaced by another, and the site calibration procedure and checking should be repeated. The above control point configuration guidelines should be followed in selecting replacement control points. CHANGES/ERRORS Specifications direct the contractor to immediately notify the engineer of any errors during staking and construction. Noted errors should be investigated as quickly as possible and may result in changes to the project model. The machine control specifications give guidance on handling model changes. It will be necessary to coordinate with the design engineer and the MDU engineer to make model changes. In cases of significant errors and changes, further consideration may have to be given to the continued use of the machine control operations on the project. Current pilot specifications provide that should the machine control technology prove to be unworkable, the engineer would pay the item of subgrade staking for the section attempted and revert back to conventional staking. 39

48 Appendix B. WisDOT Southwest Region s Revision to Specification 40

49 Construction Staking Subgrade, Item ; Construction Staking Initial Layout, Item Conform to standard spec 650 as modified in this special provision. Replace standard spec with the following: Subgrade General (1) Use global positioning system (GPS) machine guidance and verify with conventional subgrade staking on the following designated portions of the contract as follows: GPS Machine Guidance 1) Within to Conventional Subgrade Staking 1) First 1000 will require 3 stakes every 100 2) Second 1000 cl stakes every 100 and 3 stakes every 100 in and out of superelevations 3) Third 1000 cl stakes every 500 and 3 stakes every 100 in and out of superelevations 4) Remainder will evaluate when first 3000 is complete or near complete. If the first 1000 section is acceptable to the engineer then the number of stakes can be reduced in the second 1000 section and so on. If the first section is not acceptable to the engineer then the same conventional staking intervals will be required to continue. (2) The engineer may require the contractor to revert to conventional subgrade staking methods for all or part of the work at any point during construction if, in the engineer's opinion, the GPS machine guidance is producing unacceptable results. 41

50 Appendix C. Record-Keeping Forms and Examples Distributed to Pilot Project Engineers and Foremen 42

51 43

52 44

53 Appendix D. Questions and Talking Points for Second Site Visit to Hoffman Project 45

54 Talking Points for Second Site Visit To STH 57 Hoffman NOTE: This document is a list of questions and talking points for the second site visit during August, It is not a questionnaire to be filled out and returned. The questions and talking points focus upon selected, specific aspects of the specification and guidance language for the pilot project. 1. The specification allows the engineer to require conventional subgrade staking if GPS machine guidance is producing unacceptable results. In such cases, if the GPS problems are beyond the contractor s control, WisDOT will measure the additional subgrade staking for payment. Is this a good idea? Was there a need to revert to conventional staking on the pilot project? Does this part of the spec need any modification? 2. No subgrade stakes are required for work approved for GPS machine guidance. Does this part of the spec need any modification? 3. The spec requires coordination throughout the course of construction between the contractor and engineer to ensure that GPS machine guidance conforms to contract tolerances and that methods conform to the contractor s GPS work plan. This includes addressing GPS machine guidance issues at weekly progress meetings. What was your experience with this coordination on the pilot project? Does this part of the spec need any modification? 4. The spec requires provision by the contractor of a GPS rover, along with training, to the project engineer for use as needed on the project. What was your experience with this aspect of the spec on the pilot project? Does this part of the spec need any modification? 5. The spec requires submittal of a GPS work plan by the contractor to the engineer at least 10 business days before the beginning of affected grading operations. The spec goes on to describe necessary components of the GPS work plan. What was your experience with this aspect of the spec on the pilot project? Does this part of the spec need any modification? 6. The spec requires periodic sensor calibrations, checks for blade wear, and other routine adjustments. How often were these equipment checks and sensor calibrations performed? What was checked? What was calibrated? Does this part of the spec need any modification? 46

55 7. The spec requires provision of design surface data by WisDOT in LandXML format, development of the 3D model by the contractor, confirmation by the contractor that the 3D model agrees with the plans, and provision of the 3D model to the project engineer. Furthermore, errors in department-provided information are to be reported in LandXML format. WisDOT determines if revisions to the plans are required. If revisions to the plans force revisions to the 3D model, WisDOT pays for the revisions to the 3D model. The contractor is to provide the current 3D model to the engineer. We quickly learned that LandXML did not work as a data exchange format. Please describe what was done to develop and maintain the 3D model on the pilot project. Were any errors in the plans detected? Was it ever necessary to revise the plans? Did revisions to the plans force revisions to the 3D model? What are your recommendations concerning this aspect of the specification? 8. For site calibration / localization, the specification requires at least 6 horizontal and vertical control points, or two per mile, whichever is greater. Is this number sufficient? Is it over-specified? Does this part of the spec need any modification? 9. Site calibration checks are required at the start of each day and at least once for every five hours of continuous subgrade construction work. Horizontal tolerance is 0.10 ft or less. Vertical tolerance is 0.05 ft or less. Is this frequency appropriate for site calibration / localization checking? Are the tolerances appropriate? Did any site calibration / localization check fail to meet tolerance? If so, what was done? Does this part of the spec need any modification? 10. Daily site calibration / localization checks results are to be provided to the engineer at weekly progress meetings. Does this part of the spec require any modification? 11. A GPS rover is to be used by the contractor to check the subgrade at 20 or more randomly selected locations per mile. At least 4 of any 5 consecutively-tested subgrade points must be within 0.10 ft (vertically) of the plan elevation. If otherwise, the engineer must be notified. The engineer makes periodic independent subgrade checks and notifies the contractor if any individual check differs by more than 0.10 ft from design. How did the contractor select the check points? Were there any failures of the 4-out-of ft tolerance? If so, what was done? Did any of the engineer s checks fail the 0.10 ft tolerance? If so, what was done? Does this part of the spec require any modification? Can the check tolerance be tightened up, say to 0.07 ft or 0.08 ft? 47

56 12. In initial layout, the specification requires a minimum of 6 established control points per mile. These control points are to be consistent with third-order, class I horizontal and third-order vertical accuracy. Vertical control is to be established by differential leveling. Is the number of control points appropriate? Is the accuracy requirement appropriate? Is the differential leveling method appropriate? Does this part of the spec require any modification? 13. Is there any other aspect of the specification that needs attention? Are there unnecessary redundancies in the specification? Is there anything left out of the specification? What else can be done to improve the specification? 14. Have you reviewed the guidance language? Was it necessary to rely upon any of the guidance language during the pilot project? If so, did you find it useful? Even if you did not need to use the guidance language on the pilot project, do you think it is useful in its current form? Is there anything missing from the guidance language? Are there any unnecessary redundancies in the guidance language? Is there anything that is unclear or confusing in the guidance language? What else can be done to improve the guidance language? 48

57 Appendix E. Questions and Talking Points for Second Site Visit to Wondra Project 49

58 Talking Points for Second Site Visit To STH Wondra NOTE: This document is a list of questions and talking points for the second site visit during August, It is not a questionnaire to be filled out and returned. The questions and talking points focus upon selected, specific aspects of the specification and guidance language for the pilot project. 1. The specification allows the engineer to require conventional subgrade staking if GPS machine guidance is producing unacceptable results. The STH 106 spec requires some conventional subgrade staking in the first 3000 feet of the project. Was there a need to revert to conventional staking on other parts of the pilot project? If so, why? What is your recommendation for 2008 pilot projects? Should conventional subgrade staking be at the discretion of the engineer in case GPS machine guidance is not working well? Should there be any required subgrade staking (as in the 2007 STH 106 spec) or subgrade staking at the discretion of the engineer, even if GPS machine guidance is working well? 2. The spec requires coordination throughout the course of construction between the contractor and engineer to ensure that GPS machine guidance conforms to contract tolerances and that methods conform to the contractor s GPS work plan. This includes addressing GPS machine guidance issues at weekly progress meetings. What was your experience with this coordination on the pilot project? Does this part of the spec need any modification? 3. The spec requires provision by the contractor of a GPS rover, along with training, to the project engineer for use as needed on the project. What was your experience with this aspect of the spec on the pilot project? Does this part of the spec need any modification? 4. The spec requires submittal of a GPS work plan by the contractor to the engineer at least 10 business days before the beginning of affected grading operations. The spec goes on to describe necessary components of the GPS work plan. What was your experience with this aspect of the spec on the pilot project? Does this part of the spec need any modification? 5. The spec requires periodic sensor calibrations, checks for blade wear, and other routine adjustments. 50

59 How often were these equipment checks and sensor calibrations performed? What was checked? What was calibrated? Does this part of the spec need any modification? 6. The spec requires provision of design surface data by WisDOT in LandXML format, development of the 3D model by the contractor, confirmation by the contractor that the 3D model agrees with the plans, and provision of the 3D model to the project engineer. Furthermore, errors in department-provided information are to be reported in LandXML format. WisDOT determines if revisions to the plans are required. If revisions to the plans force revisions to the 3D model, WisDOT pays for the revisions to the 3D model. The contractor is to provide the current 3D model to the engineer. We quickly learned that LandXML did not work as a data exchange format. Please describe what was done to develop and maintain the 3D model on the pilot project. Were any errors in the plans detected? Was it ever necessary to revise the plans? Did revisions to the plans force revisions to the 3D model? What are your recommendations concerning this aspect of the specification? 7. For site calibration / localization, the specification requires at least 6 horizontal and vertical control points, or two per mile, whichever is greater. Is this number sufficient? Is it over-specified? Does this part of the spec need any modification? 8. Site calibration checks are required at the start of each day and at least once for every five hours of continuous subgrade construction work. Horizontal tolerance is 0.10 ft or less. Vertical tolerance is 0.05 ft or less. Is this frequency appropriate for site calibration / localization checking? Are the tolerances appropriate? Did any site calibration / localization check fail to meet tolerance? If so, what was done? Does this part of the spec need any modification? 9. The site calibration / localization file is to be provided to the project engineer. Daily site calibration / localization checks results are to be provided to the engineer at weekly progress meetings. Does this part of the spec require any modification? 10. A GPS rover is to be used by the contractor to check the subgrade at 20 or more randomly selected locations per mile. At least 4 of any 5 consecutively-tested subgrade points must be within 0.10 ft (vertically) of the plan elevation. If otherwise, the engineer must be notified. The engineer makes periodic independent subgrade checks and notifies the contractor if any individual check differs by more than 0.10 ft from design. 51

60 How did the contractor select the check points? Were there any failures of the 4-out-of ft tolerance? If so, what was done? Did any of the engineer s checks fail the 0.10 ft tolerance? If so, what was done? Does this part of the spec require any modification? Can the check tolerance be tightened up, say to 0.07 ft or 0.08 ft? 11. Is there any other aspect of the specification that needs attention? Are there unnecessary redundancies in the specification? Is there anything left out of the specification? What else can be done to improve the specification? 12. Have you reviewed the guidance language? Was it necessary to rely upon any of the guidance language during the pilot project? If so, did you find it useful? Even if you did not need to use the guidance language on the pilot project, do you think it is useful in its current form? Is there anything missing from the guidance language? Are there any unnecessary redundancies in the guidance language? Is there anything that is unclear or confusing in the guidance language? What else can be done to improve the guidance language? 52

61 Appendix F. GPS Work Plan Submitted by Hoffman Construction 53

62 54

63 55

64 56

65 57

66 58

67 Appendix G. Site Calibration / Localization Checks for Hoffman Project 59

68 Two Pages of Calibration Checks 60

69 61

70 Appendix H. Subgrade Checks for Hoffman Project 62

71 63

72 64

73 65

74 66

75 67